In most practical investigations, inhomogeneous backgrounds and noise mask the objects in IRT-recorded thermal images. Although IRT offers several advantages in terms of detection efficiency, its direct performance does not provide satisfactory results. Infrared thermography (IRT), as a classical non-destructive test technique, has gained wide interest in CFRP quality assessment and cultural heritage restoration. The unknown sizes, shapes, locations, and physical properties of the defects make the NDE study of CFRP a challenge. During the manufacturing and long-term service of CFRP products, defects are inevitably generated inside the materials. Research on its nondestructive evaluation (NDE) and structural health monitoring has become a necessary and interesting topic. Experimental results on carbon fiber-reinforced polymers demonstrate the superiorities of GMLT, compared with other methods.Īs one of the advanced composite materials, the demand for carbon fiber-reinforced polymer (CFRP) in new energy, equipment manufacturing, and other fields is growing rapidly. Moreover, probability density maps and quantitative metrics are proposed to evaluate and explain the obtained defect detection performance. Finally, the partial least squares regression is presented to extract the explicit mapping of manifold learning for defect visualization. Subsequently, the manifold learning method is employed for the unsupervised dimensionality reduction in all images. Specifically, the spectral normalized generative adversarial networks serve as an image augmentation strategy to learn the thermal image distribution, thereby generating virtual images to enrich the dataset. In this work, a novel generative manifold learning thermography (GMLT) is proposed for defect detection and the evaluation of composites. However, the performance of these methods is still restricted by limited informative images and difficulties in feature extraction caused by inhomogeneous backgrounds and noise. Infrared thermography techniques with thermographic data analysis have been widely applied to non-destructive tests and evaluations of subsurface defects in practical composite materials.
0 Comments
Leave a Reply. |